Equipment Status Monitoring System

Abstract

To make it possible to detect an abnormality in a plurality of facilities to be monitored without equipping each facility with a vibration sensor or the like.SOLUTION: A facility condition monitoring system comprises: a sensor 11 that outputs data for indicating a condition of a facility 2 as sensor data which is used for condition monitoring; a communication unit 13 that transmits the sensor data; a power supply unit 12 that supplies electric power to the sensor and the communication unit; and a common sensor node 10 provided for a plurality of facilities. Also, the facility condition monitoring system comprises: a receiver 20 that receives the sensor data transmitted from the communication unit; and a condition detection unit 30 that inputs the sensor data received by the receiver, and learns normal conditions of the plurality of facilities as learning data based on the sensor data in a normal condition when the plurality of facilities are operating normally, the condition detection unit comparing conditions of plural facilities indicated by the sensor data with the learning data to detect occurrence of abnormalities or signs of abnormalities in the plural facilities when the sensor data transmitted from the sensor node after learning is received by the receiver.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a facility status monitoring system that monitors facility abnormalities. [Background technology]

[0002] Conventionally, inspection devices for detecting abnormalities in operating equipment have been known. For example, Patent Document 1 proposes an inspection device for determining the state of a rotating device, which includes a vibration sensor, a correlation diagram generation unit, a deep learning unit, and a judgment unit. When the vibration state of bearings in an operating rotating device is detected by multiple vibration sensors, the correlation diagram generation unit generates a correlation diagram showing the correlation between multiple acceleration signals output by the multiple vibration sensors. Then, the deep learning unit performs deep learning based on the correlation diagram generated by the correlation diagram generation unit, and the judgment unit judges the state of the rotating part based on the results of the deep learning, making it possible to detect abnormalities in various devices with different configurations. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-091640 Summary of the Invention [Problem to be solved by the invention]

[0004] However, with conventional inspection equipment, in an environment with many operating machines, such as a production line, a large number of vibration sensors are required to be able to detect abnormalities in a large number of operating machines.Furthermore, given that equipment abnormalities are caused by multiple factors, multiple types of sensing are required to detect the details of the abnormality, such as the abnormal location and cause, which requires a huge number of sensors, making real-time monitoring difficult.

[0005] In view of the above, it is an object of the present invention to provide a facility status monitoring system that can detect abnormalities in multiple monitored objects without providing a vibration sensor or the like at each monitored object. [Means for solving the problem]

[0006] In order to achieve the above object, the invention described in claim 1 is an equipment status monitoring system, comprising: a sensor node (10) that has a sensor (11) that outputs data indicating the status of equipment (2) to be monitored as sensor data, a communication unit (13) that transmits the sensor data, and a power supply unit (12) that supplies power to the sensor and the communication unit, and that is commonly applied to multiple monitored objects; a receiver (20) that receives the sensor data transmitted from the communication unit; and a status detection unit (30) that inputs the sensor data received by the receiver and learns the normal status of the multiple monitored objects as learning data based on the sensor data in normal conditions when the multiple monitored objects are operating normally, and when the sensor data transmitted from the sensor node is received by the receiver after learning, compares the status of the multiple monitored objects indicated by the sensor data with the learning data to detect the occurrence of an abnormality or a sign of an abnormality in the multiple monitored objects, wherein the method of power supply from the power supply unit to the sensor and communication unit is one of a method using a magnetic field, energy harvesting, and mobile battery power supply, and the sensor node is placed on a mobile object and moved together with the mobile object to acquire sensor data indicating the status of the multiple monitored objects.

[0007] In this way, for multiple monitored objects, at least one common sensor node is used to transmit sensor data when the multiple monitored objects are operating normally to the state detection unit, and the sensor data is learned as learning data for the normal state. Therefore, by comparing the states of the multiple monitored objects indicated by the sensor data transmitted from the sensor node after learning with the learning data, it becomes possible to detect abnormalities in the multiple monitored objects without providing each monitored object with a vibration sensor or the like.

[0008] The reference symbols in parentheses attached to each component indicate an example of the correspondence between the component and the specific components described in the embodiments described below. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram of an equipment state monitoring system according to a first embodiment. [Figure 2A] This is a schematic diagram showing two microphones arranged so that their directivities are offset by 90°. [Figure 2B] FIG. 2B is a diagram showing the polar pattern when two microphones are arranged as shown in FIG. 2A. [Figure 3A] FIG. 2B is a diagram showing a state in which a sound source is placed midway between the x-axis and y-axis when two microphones are placed as in FIG. 2A. [Figure 3B] 3B is a diagram showing the received sound pressure at each microphone when sound is output from the sound source shown in FIG. 3A. FIG. [Figure 4A] FIG. 2B is a diagram showing a state in which a sound source is arranged on the x-axis when two microphones are arranged as in FIG. 2A. [Figure 4B] 4B is a diagram showing the received sound pressure at each microphone when sound is output from the sound source shown in FIG. 4A. FIG. [Figure 5] FIG. 10 is a diagram showing a case where sensor nodes are arranged so as to flow on three transport paths. [Figure 6] FIG. 2 is a transparent perspective view of a sensor node. [Figure 7] FIG. 2 is an exploded view of a sensor node. [Figure 8] FIG. 10 is a diagram illustrating a case where the center of gravity of the sensor node is located above the center in the vertical direction. [Figure 9] FIG. 10 is a diagram illustrating an example in which a sensor node has a vibration suppression structure. [Figure 10] FIG. 10 is a diagram showing a vibration suppression structure in the form of a through hole. [Figure 11] FIG. 10 is a diagram showing an example in which a workpiece directly placed on a conveying path is replaced with a sensor node. [Figure 12]FIG. 10 is a diagram showing an example in which workpieces arranged on a pallet are replaced with sensor nodes. [Figure 13] FIG. 10 is a diagram showing an example in which a workpiece placed on a jig is replaced with a sensor node. [Figure 14] FIG. 10 is a diagram showing an example of an arrangement in which a sensor node is directly attached to a workpiece. [Figure 15] FIG. 10 is a diagram showing an example of an arrangement in which a sensor node is directly attached to a workpiece. [Figure 16] FIG. 10 is a diagram showing an example of an arrangement in which a sensor node is directly attached to a pallet. [Figure 17] FIG. 10 is a diagram showing an example in which a sensor node is arranged above the center of a workpiece in the vertical direction. [Figure 18] FIG. 10 is a diagram showing an example in which a sensor node is arranged on the rear side of a workpiece in the direction of travel. [Figure 19] FIG. 1 is a diagram showing a configuration in which an equipment status monitoring system includes a server. [Figure 20] FIG. 10 is a diagram showing an example of history information when the equipment status monitoring system functions as a traceability system. [Figure 21] FIG. 10 is a diagram showing a state in which a plurality of sensor nodes are arranged on a transport path. [Figure 22] FIG. 10 is a diagram showing another example of the configuration of the composite sensor. [Figure 23] FIG. 10 is a diagram showing another example of the configuration of the composite sensor. [Figure 24A] FIG. 2 is a diagram showing an example of the configuration of one sensor included in the composite sensor. [Figure 24B] FIG. 10B is a transparent perspective view of a composite sensor configured using sensors having the configuration of FIG. 10A. [Figure 25] FIG. 10 is a diagram showing a state in which a sensor node is installed on a transport path. [Figure 26] FIG. 10 is a diagram showing frequency characteristics when sound pressure is measured by a sensor node. [Figure 27] 10A and 10B are diagrams for explaining a method for identifying the position of a transport path that is divided into a plurality of sections. [Figure 28] FIG. 2 is a diagram showing details of a state detection unit, such as functional blocks. [Figure 29] FIG. 2 is a diagram showing how each piece of equipment is arranged along a conveying path. [Figure 30] FIG. 30 is a diagram showing the degree of abnormality at each position when the sensor node moves on the transport path shown in FIG. 29. [Figure 31] FIG. 10 is a diagram showing an example of displaying detection results on a display device. [Figure 32] 10A and 10B are diagrams for explaining a method for detecting abnormalities occurring in a plurality of transport paths. [Figure 33] 10A and 10B are diagrams illustrating an example of a detection result of an abnormality occurring in a transport path displayed on a display device. [Figure 34] FIG. 10 is a block diagram illustrating a process of placing an order with a parts manufacturer based on the detection result of a state detection unit. [Figure 35] 1 is a graph showing an example of the relationship between energy consumption and production amount. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following embodiments, parts that are identical or equivalent to each other will be denoted by the same reference numerals.

[0011] (First embodiment) A first embodiment will be described below. The equipment state monitoring system of this embodiment monitors abnormalities in a plurality of pieces of equipment to be monitored by using a common sensor node equipped with a sensor.

[0012] As shown in Fig. 1, the equipment status monitoring system 1 is configured to include a sensor node 10 that monitors the status of equipment 2 to be monitored, a receiving unit 20, a status detection unit 30, and a display device 40. The monitoring results are displayed on the display device 40, and, for example, an operator 3 can control, repair, or replace parts of the equipment 2 based on the displayed content, thereby maintaining the equipment 2 in good condition. Note that although Fig. 1 shows only one equipment 2, there may be multiple equipment 2, and the number is arbitrary.

[0013] [Configuration of the sensor node 10] The sensor node 10 is equipped with at least one sensor 11 to monitor abnormalities in multiple pieces of equipment 2, and is configured to include, in addition to the sensor 11, a power supply unit 12, a communication unit 13, etc.

[0014] The sensor 11 detects one or more of the following as detection targets: sound, vibration or acceleration, angular velocity, temperature, humidity, magnetism, light, surrounding images, flow rate, pressure, odor, etc. When detecting more than one of these, or when detecting more than one of the same detection targets, a composite sensor is provided with multiple sensors 11. The sensor 11 may be configured with any type of sensor, such as a semiconductor sensor, and outputs sensing signals indicating the detection results as sensor data to the communication unit 13. The sensor data from this sensor 11 corresponds to data indicating various states, such as physical quantities, used to monitor the state of the equipment 2.

[0015] The sensor 11 may be a sensor corresponding to the above-mentioned detection target. For example, a sound sensor may be used to detect sound, a vibration sensor to detect vibration or acceleration, an angular velocity sensor to detect angular velocity, a temperature and humidity sensor to detect the temperature and humidity of the surrounding atmosphere, or a flow sensor to detect the surrounding air volume. The sensor 11 may also be a magnetic sensor to detect magnetism, an optical sensor to detect light, or an image sensor composed of a camera or the like to detect a surrounding image. The sensor 11 may also be a flow sensor to detect flow rate, a pressure sensor to detect pressure, or an odor sensor to detect odor. As described above, the sensor 11 may be of multiple types with different detection targets, or may include multiple sensors with the same detection target. Even if the multiple sensors 11 detect the same target, if the detection directions or positions are different, sensor data corresponding to the directions or positions can be obtained.

[0016] In the case of a sound sensor, multiple microphones may be used to identify the position of a sound source. Direction can be determined using phase difference or time difference, or by pointing multiple microphones with the same polar pattern (i.e., directionality) in different directions and utilizing the sound pressure difference resulting from differences in sensitivity. For example, if one direction of the sensor node 10 is defined as the front, the opposite direction as the rear, and the directions on either side of that as the left and right directions, then four unidirectional microphones can be oriented in four directions (front, back, left, right, front, and rear). In this way, for example, for sound input from the left, differences in input sound pressure due to differences in sensitivity occur, such that the left microphone experiences high sound pressure, the front and rear microphones experience medium sound pressure, and the right microphone experiences low sound pressure. This allows the direction of a sound source to be determined based on a previously determined polar pattern.

[0017] For example, when using two microphones to identify the direction of a sound source, two microphones 11a and 11b are positioned with their directivities offset by 90°, as shown in FIG. 2A. In this case, as shown in FIG. 2B, the polar pattern 11c of microphone 11a has a pattern in which the reference sound pressure is 0 dB at the 0° position, i.e., the position in the positive direction of the y-axis, and the sound pressure gradually decreases toward the 90° position, i.e., the position in the positive direction of the x-axis. Conversely, the polar pattern 11d of microphone 11b has a pattern in which the reference sound pressure is 0 dB at the 90° position, i.e., the position in the positive direction of the x-axis, and the sound pressure gradually decreases toward the 0° position, i.e., the position in the positive direction of the y-axis. In this case, if sound source 4 is located at a 45° angle, midway between the x-axis and the y-axis, as shown in FIG. 3A, the received sound pressures of microphones 11a and 11b will be approximately the same, as shown in FIG. 3B. Furthermore, for example, if sound source 4 is located at an angle of 90° on the x-axis as shown in FIG. 4A, the sound pressure received by microphone 11b will be about twice the sound pressure received by microphone 11a as shown in FIG. 4B. In this way, by comparing the sound pressure received by multiple microphones 11a and 11b arranged with shifted directivity, the direction of sound source 4 can be identified. The distance to sound source 4 can also be determined from the sound pressure. Therefore, the direction and distance to sound source 4 can be identified by a sound sensor using multiple microphones 11a and 11b.

[0018] Furthermore, a vibration sensor can detect rattles in the conveyance path, which cannot be detected by a sound sensor. A vibration sensor can detect vibrations and acceleration of the detection target. The vibration sensor can then detect abnormalities in the detection target based on the vibrations and acceleration. Furthermore, if the sensor node 10 is installed on the conveyance path, the vibration sensor can detect vibrations in the conveyance path, but if the sensor 11 is equipped with both a vibration sensor and a sound sensor, it can be determined whether the vibrations are originating from the conveyance path. In other words, if an abnormality is detected based on the data of the detection results acquired by both the vibration sensor and the sound sensor or the data acquired only by the vibration sensor, it can be determined that the abnormality is originating from the conveyance path, whereas if an abnormality is detected only by the sound sensor, it can be determined that the abnormality is not originating from the conveyance path but is due to some other factor. This makes it possible to more clearly distinguish whether the abnormality is due to the conveyance path or some other factor, and identify the equipment 2 in which the abnormality was detected.

[0019] Furthermore, in the case of an angular velocity sensor, changes in the rotation and orientation of the monitored object can be detected as angular velocity. An abnormality in the monitored object can be detected based on the rotation and orientation of the monitored object detected by the angular velocity sensor. Furthermore, if the sensor node 10 is placed, for example, on a conveyance path along which a product is conveyed, or on a product conveyed along the conveyance path, the angular velocity sensor can detect changes in the inclination and orientation of the conveyance path and the product.

[0020] Furthermore, the angular velocity sensor can be at least one of a 1-axis angular velocity sensor that detects rotation in one direction, a 2-axis angular velocity sensor that detects rotation in two mutually orthogonal directions, and a 3-axis angular velocity sensor that detects rotation in three mutually orthogonal directions.

[0021] In addition, in the case of an optical sensor, multiple light receiving units may be used to identify the position of the light source. The method of direction identification may be to use the phase difference or time difference of light, or to orient multiple light receiving units in different directions and identify the direction using the difference in the amount of light received. For example, the sensor node 10 may have light receiving units facing four directions: forward, backward, and left and right. In this way, for example, for light input from the left, differences in the amount of light received will occur, such that the left light receiving unit receives a large amount of light, the front and rear light receiving units receive a medium amount of light, and the right light receiving unit receives a small amount of light. Therefore, the direction of the light source can be identified based on the difference in the amount of light received.

[0022] The power supply unit 12 is a power supply source for each component of the sensor node 10, and is configured using, for example, a button battery or a lithium battery. The power supply unit 12 can be configured in a variety of ways: by using a magnetic field, such as an electromagnetic induction system, a magnetic field resonance system, an electric field coupling system, or a radio wave reception system; by harvesting energy from vibration, light, heat, or electromagnetic waves; or by using a mobile battery. Depending on how the sensor node 10 is used, for example, if it is used as a carrier, the power supply unit 12 can be configured by selecting the optimal power supply system based on the power, transmission distance, size, and other factors required for the carrier.

[0023] The communication unit 13 transmits the sensor data transmitted from the sensor 11 to the receiving unit 20 by wireless communication or the like. The frequency band to be used is selected based on the communication speed, communication distance, optimal frequency wraparound characteristics, etc. of the communication unit 13. A microcomputer may also be installed in the communication unit 13 to perform various controls based on the sensing signal of the sensor 11. For example, if the sensor 11 is a sound sensor, the communication unit 13 may perform control to determine whether or not to transmit the sensor data, such as by communicating only when the sound pressure at a predetermined frequency exceeds a preset threshold.

[0024] It is also preferable to extend the operating time of the power supply unit 12 even with the same battery capacity. For this reason, rather than constantly performing data communication, it is preferable to control the communication unit 13 so that only the data and the data before and after it are transmitted when a preset threshold is exceeded by the microcomputer of the communication unit 13. For example, the amount of data transmitted as the data before and after the threshold can be appropriately set to, for example, 10 cycles before and after the threshold. This reduces the amount of communication and improves the battery life of the power supply unit 12.

[0025] Furthermore, when the sensor data transmitted from the sensor 11 to the communication unit 13 is the sensor data detected by the sensor 11 as is, that is, when the sensor data received by the communication unit 13 is a raw value, the communication unit 13 may process the sensor data and transmit it to the receiving unit 20. An example of when the communication unit 13 processes and transmits the sensor data will be described with reference to FIG. 5.

[0026] 5, a sensor node 10 having a vibration sensor is arranged to flow on three transport paths 5a, 5b, and 5c. The three transport paths, the first transport path 5a, the second transport path 5b, and the third transport path 5c, are connected in this order and move at different transport speeds, thereby transporting the sensor node 10 in synchronization with the transport speed at which the sensor node 10 itself operates. In this case, the communication unit 13 may transmit the sensor data received from the sensor 11 to the receiving unit 20 in accordance with the transport speeds of the first transport path 5a, the second transport path 5b, and the third transport path 5c, respectively.

[0027] For example, let us assume that the conveying speed of the first conveying path 5a is a first speed v1, the conveying speed of the second conveying path 5b is a second speed v2, and the conveying speed of the third conveying path 5c is a third speed v3, with the first speed v1 being slower than the second speed v2 and the third speed v3. In this case, the vibrations generated when the first conveying path 5a operates have a slower vibration period than the vibrations generated when the second conveying path 5b and the third conveying path 5c operate. Therefore, the sampling frequency required for the vibration sensor to detect an abnormality or a sign of an abnormality in the first conveying path 5a can be lower than the sampling frequency required to detect an abnormality or a sign of an abnormality in the second conveying path 5b and the third conveying path 5c.

[0028] Therefore, the communication unit 13 does not need to transmit raw values ​​of sensor data detected at the same sampling frequency on all of the first conveying path 5a, the second conveying path 5b, and the third conveying path 5c to the receiving unit 20. For example, the communication unit 13 may transmit sensor data that has been downsampled from raw values ​​of sensor data detected on the first conveying path 5a, which has a slower conveying speed than the second conveying path 5b and the third conveying path 5c, to the receiving unit 20. In this way, the amount of data transmitted from the communication unit 13 to the receiving unit 20 is reduced, which reduces the amount of communication and improves the battery life of the power supply unit 12.

[0029] The sampling frequency of the downsampling process may be set in advance by the operator 3 according to the respective conveying speeds of the first conveying path 5a, the second conveying path 5b, and the third conveying path 5c. Alternatively, the sampling frequency of the downsampling process may be determined by the communication unit 13 based on a control map that is set in advance in the communication unit 13 based on the respective conveying speeds and sampling frequencies of the first conveying path 5a, the second conveying path 5b, and the third conveying path 5c. Furthermore, the sampling frequency of the downsampling process may be determined by the communication unit 13 based on the conveying speeds of each conveying path detected by the detection operation, using the sensor node 10. In this case, the communication unit 13 performs downsampling at a sampling frequency for a conveying path having a slower conveying speed than the sampling frequency for a conveying path having a faster conveying speed.

[0030] [Configuration example of sensor node 10] Next, an example of the configuration of the sensor node 10 will be described with reference to FIGS.

[0031] FIG. 6 shows the overall configuration of a sensor node 10 in which the sensor 11 is a composite sensor. As shown in this figure, the sensor 11, power supply unit 12, and communication unit 13 are integrated into a polyhedron, here a hexahedron. This integrated structure is then housed within a hexahedron-shaped housing 14, abutting against the inner wall of the housing 14 to prevent any gaps, thereby firmly securing the sensor node 10. The housing 14 is made of a material appropriate for the environment in which it will be used. For example, if the sensor node 10 will be used in an environment where it will be exposed to water, it will be designed with a water-resistant material to protect the sensor 11 and other components. Note that in FIG. 6, the housing 14 is shown as a see-through view so that the integrated structure of the sensor 11, power supply unit 12, and communication unit 13 arranged within the housing 14 can be seen.

[0032] Incidentally, when the sensor node 10 has a vibration sensor for detecting vibrations of the conveyance path 5, the sensor node 10 detects vibrations of the conveyance path 5 caused by the influence of the operation of the conveyance path 5. However, vibrations caused by influences other than the operation of the conveyance path 5 can become noise in the vibrations of the conveyance path 5 detected by the vibration sensor.

[0033] 8, for example, assume that the center of gravity Cg of the sensor node 10 is located above the vertical center of the sensor node 10. In this case, if wind flowing toward the sensor node 10 hits the sensor node 10, there is a risk that the sensor node 10 itself will vibrate due to this wind. If the sensor node 10 vibrates due to the impact of the wind and the sensor node 10 detects its own vibration, which is vibration other than that of the transport path 5, the vibration of the sensor node 10 itself will become noise in addition to the vibration of the transport path 5 detected by the sensor node 10.

[0034] Furthermore, if the sensor node 10 has a sound sensor for detecting sounds around the sensor node 10, the sensor node 10 detects changes in sound pressure in the surrounding environment of the sensor node 10 as air vibrations caused by the sound. However, vibrations caused in the air by factors other than sound pressure can become noise in the air vibrations detected by the sound sensor. For example, if wind flows in the surrounding environment of the sensor node 10 and vibrates the air in the surrounding environment of the sensor node 10, the vibrations caused by this wind become noise in the air vibrations detected as sound by the sensor node 10. In addition, sound caused by wind flowing toward the sensor node 10 hitting the sensor node 10 also becomes noise in the sounds in the surrounding environment of the sensor node 10 detected by the sensor node 10.

[0035] The wind that causes these vibration noises includes, for example, natural wind flowing in the surrounding environment of the conveying path 5, and vertical laminar flows generated by an air blowing process installed along the conveying path 5 or a blower installed in a clean room, etc.

[0036] In contrast to this, the sensor node 10 may have a vibration suppression structure that suppresses vibration of the sensor node 10 due to factors other than vibration of the transport path 5, even when there is wind flowing toward the sensor node 10. This vibration suppression structure may be a structure that offsets the position of the center of gravity Cg of the sensor node 10 downward from the center of the sensor node 10 in the vertical direction, as shown in FIG.

[0037] Specifically, the vibration suppression structure may be a structure in which the power supply unit 12, which is a relatively heavy component in the sensor node 10, is disposed below the vertical center of the sensor node 10. This allows the center of gravity Cg of the sensor node 10 to be offset below the vertical center.

[0038] Furthermore, when the housing 14 is made of a plurality of materials with different masses per unit volume, the vibration suppression structure may be configured such that the lower side in the vertical direction of the housing 14 is made of a material with a larger mass per unit volume and the upper side is made of a material with a smaller mass per unit volume. This allows the center of gravity Cg of the sensor node 10 to be offset downward from the center in the vertical direction.

[0039] Furthermore, although not shown, the vibration suppression structure may be a structure in which the center of gravity Cg of the sensor node 10 is offset vertically downward by making the size of the housing 14 larger on the lower side than on the upper side in the vertical direction. Also, the vibration suppression structure may be a structure in which the center of gravity Cg of the sensor node 10 is offset vertically downward by attaching a weight member that offsets the center of gravity Cg of the sensor node 10 below the vertical center of the housing 14 of the sensor node 10.

[0040] By providing the sensor node 10 with such a vibration suppression structure, it is possible to improve the stability of the sensor node 10 when it is placed on the transport path 5, even when there is wind flowing toward the sensor node 10. Therefore, it is possible to reduce vibrations of the sensor node 10 itself that are caused by wind flowing toward the sensor node 10 hitting the sensor node 10.

[0041] 10, the vibration suppression structure may be a fluid design structure that is less susceptible to the influence of wind even when wind is flowing toward the sensor node 10. Specifically, when wind is flowing toward the sensor node 10, the vibration suppression structure may be configured with through-holes 141 that are formed so as to penetrate in a direction corresponding to the direction of the wind flowing toward the sensor node 10.

[0042] For example, if there is wind flowing from the front side to the rear side of the sensor node 10 in the direction of travel of the transport path 5, the through hole 141 may be formed to penetrate along the direction of travel of the transport path 5, which is the direction along which this wind flows.

[0043] By configuring the sensor node 10 to have such a through hole 141, even if there is wind flowing toward the sensor node 10, the through hole 141 can deflect the wind from the front side to the rear side in the traveling direction of the transport path 5. Therefore, it is possible to reduce vibration of the sensor node 10 itself, which is caused when wind flowing toward the sensor node 10 collides with the sensor node 10.

[0044] Furthermore, by allowing the wind to flow through the through-holes 141, it is possible to suppress the generation of noise caused by the wind hitting the sensor node 10. Therefore, even if the sound sensor is provided on the rear side or side of the housing 14 in the direction of travel, it is possible to make it easier for the sound sensor to detect sounds in the environment surrounding the sensor node 10.

[0045] Note that the through-holes 141 may be inclined upward or downward, or left or right, from the front to the rear in the direction of travel of the transport path 5, as long as they can deflect the wind flowing from the front to the rear in the direction of travel of the transport path 5. Furthermore, the fluid design structure that makes the sensor node 10 less susceptible to the influence of the wind flowing toward it may be different from the configuration in which the through-holes 141 are formed in the sensor node 10. Although not shown, for example, the shape of the housing 14 of the sensor node 10 may be such that the cross-sectional area of ​​the housing 14 perpendicular to the direction of travel of the transport path 5 decreases from the rear to the front in the direction of travel of the transport path 5, thereby making it possible to make the sensor node 10 less susceptible to the influence of the wind flowing toward it.

[0046] 5 and other examples, the integrated structure of the sensor 11, power supply unit 12, and communication unit 13 is hexahedral in shape, but the hexahedral shape makes it easier to set the orientation of the sensor node 10, i.e., the orientation of the sensor 11. For this reason, the integrated structure does not necessarily have to be hexahedral in shape, and may be another polyhedral shape.

[0047] The sensor node 10 can monitor multiple pieces of equipment 2 even when placed in a specific location, and if it is configured as a composite sensor, it can monitor multiple pieces of equipment 2 in more detail. However, if it is installed on a mobile object, it can be moved to monitor multiple pieces of equipment 2. For example, when the sensor node 10 is used to monitor production equipment, it can be installed on a transport path that serves as a mobile object and moved along with the transport path. For example, by installing the sensor node 10 on a belt conveyor that serves as the transport path, the sensor node 10 can be transported together with the work on the belt conveyor. Furthermore, the sensor node 10 may be installed on a mobile object such as an AGV (Automatic Guided Vehicle) and moved along with the mobile object to monitor the status of surrounding equipment 2.

[0048] The sensor node 10 may be assembled in any manner, but if the sensors 11 include a vibration sensor or sound sensor, it is preferable to fix the sensor node 10 to the installation location using as strong a method as possible, such as welding or screwing, in order to reduce noise in the acquired sensor signal. Of course, depending on the installation location, the sensor node 10 may also be installed using a magnet or adhesive.

[0049] 11 to 18, a specific example of a method for arranging the sensor nodes 10 on the transport path 5 together with a plurality of workpieces W that are arranged side by side on the transport path 5, which is a moving body, and that are transported by the transport path 5, will be described. The workpieces W may be, for example, objects to be processed by processing equipment, or may be products that are being manufactured or have been manufactured.

[0050] The sensor node 10 placed on the transport path 5 is transported by the transport path 5 together with multiple workpieces W. When the sensor node 10 is placed on the transport path 5, there are cases where one of the multiple workpieces W is replaced with the sensor node 10, and cases where the sensor node 10 is attached to one of the multiple workpieces W.

[0051] 11 to 13 show examples in which the sensor node 10 replaces the workpiece W and the sensor node 10 and the workpiece W are spaced apart from each other. In these three configurations in which the workpiece W is replaced by the sensor node 10, the sensor node 10 becomes the transport body in place of the workpiece W, which is the original transport body on the transport path 5.

[0052] 14 in which the sensor node 10 is attached to the workpiece W placed directly on the conveying path 5, and two forms shown in FIGS. 15 and 16 in which the sensor node 10 is attached to the workpiece W placed on the conveying path 5 via a pallet P. In these three forms in which the sensor node 10 is attached to the workpiece W, the sensor node 10 also becomes a conveyed body together with the workpiece W, which is a conveyed body on the conveying path 5.

[0053] Of the three modes in which the sensor node 10 and the workpiece W are transported while being spaced apart from each other, the first mode shown in Fig. 11 is adopted when the workpiece W is placed directly on the transport path 5. The sensor node 10 is then placed on the transport path 5 and transported as a transport body.

[0054] When such a workpiece W is placed directly on the conveying path 5, the workpiece W may be moved from the conveying path 5 to the processing equipment to be processed by the processing equipment. In this case, the workpiece W is moved to the processing equipment by being clamped and lifted by chuck equipment (not shown) arranged around the conveying path 5, for example.

[0055] In contrast, when the sensor node 10 is placed in place of the workpiece W as in the first embodiment shown in FIG. 11 , the sensor node 10 is clamped by the chucking equipment in the same manner as the workpiece W, thereby making it possible to monitor the chucking equipment. For example, when the sensor node 10 is configured to have a pressure sensor at the location clamped by the chucking equipment, the pressure sensor detects the pressure when the sensor node 10 is clamped, thereby making it possible to monitor whether the chucking equipment is operating normally. Furthermore, when the sensor node 10 is configured to have an angular velocity sensor, the posture of the sensor node 10 when clamped and lifted by the chucking equipment can be detected, thereby making it possible to monitor whether the chucking equipment is operating normally.

[0056] Of the three modes in which the sensor node 10 and the workpiece W are transported while being spaced apart from each other, the second mode shown in FIG. 12 is adopted when the workpiece W is placed on a pallet P and transported, and the shapes of the workpiece W and the sensor node 10 are the same. In this case, the sensor node 10 is replaced with one of the multiple works W placed on the pallet P. Then, the workpiece W and the sensor node 10 are placed on the transport path 5 via the pallet P and transported as a transport body. Note that the second mode shown in FIG. 12 may also be applied when the shapes of the part of the workpiece W fixed to the pallet P and the part of the sensor node 10 fixed to the pallet P are the same.

[0057] Of the three modes in which the sensor node 10 and the workpiece W are transported while being spaced apart from each other, the third mode shown in Fig. 13 is adopted when the workpiece W is placed on a pallet P and transported, and the shapes of the workpiece W and the sensor node 10 are different from each other. In this case, the sensor node 10 replaces one of the multiple works W placed on the pallet P, and is placed together with a jig J for fixing the sensor node 10 itself to the pallet P. The sensor node 10 is then placed on the transport path 5 via the pallet P and the jig J, and transported as a transport body.

[0058] 14 shows a form in which the sensor node 10 is attached to a workpiece W placed directly on the transport path 5. Two forms in which the sensor node 10 is attached to a workpiece W placed on the transport path 5 via a pallet P are shown in FIGS. 15 and 16. When the sensor node 10 is attached to a workpiece W placed on the transport path 5 via a pallet P, there are two cases in which the sensor node 10 is attached directly to the workpiece W, and the sensor node 10 is attached to the workpiece W via the pallet P. The sensor node 10 is then placed on the transport path 5 together with the workpiece W as a transport body and transported.

[0059] In this way, when the sensor node 10 is transported as a transport body together with the workpiece W, when the workpiece W is processed by the processing equipment, the workpiece W with the sensor node 10 attached can also be processed. Furthermore, by the sensor node 10 detecting the processing state when the workpiece W is processed, it is possible to monitor whether the processing equipment is operating normally. For example, if the sensor node 10 has a vibration sensor, the sensor node 10 can detect vibrations when the processing equipment processes the workpiece W, thereby making it possible to monitor whether the processing equipment is operating normally.

[0060] Incidentally, when detecting vibrations of the transport path 5 in order to monitor the transport path 5, it is conceivable to configure the sensor node 10, which is transported together with the workpiece W as a transport body, with a vibration sensor and place it on the transport path 5. When the sensor node 10 and the workpiece W are placed on the transport path 5, if the transport body itself, which is an integrated body of the sensor node 10 and the workpiece W, vibrates due to a factor other than the vibration of the transport path 5, this may become noise in the vibration of the transport path 5 detected by the vibration sensor.

[0061] In contrast, the sensor node 10 may be attached so that the position of the center of gravity of the transport body formed by the sensor node 10 and the workpiece W is offset downward in the vertical direction. For example, when the sensor node 10 is attached to the workpiece W that is placed directly on the transport path 5, the sensor node 10 may be placed below the vertical center of the workpiece W, as shown in FIG. 14. Furthermore, when the sensor node 10 is attached to the workpiece W that is placed on the transport path 5 via a pallet P, the sensor node 10 may be attached to the pallet P instead of the workpiece W, as shown in FIG. 16.

[0062] This makes it possible to make it difficult for the transport body, on which the sensor node 10 and the work W are integrated, to vibrate due to factors other than the vibration of the transport path 5. Therefore, it is possible to suppress vibration noise caused by the vibration of the transport body itself, on which the sensor node 10 and the work W are integrated.

[0063] Furthermore, when the sensor node 10 attached to the workpiece W or the pallet P detects vibrations of the conveyance path 5, the sensor node 10 may be attached in a location where it is easy to detect vibrations of the conveyance path 5. Specifically, the sensor node 10 may be attached in a location where the amount of vibration in the workpiece W or the pallet P is likely to increase when the workpiece W or the pallet P vibrates together with the conveyance path 5 due to vibrations of the conveyance path 5. Here, when the workpiece W or the pallet P vibrates together with the conveyance path 5, the amount of vibration of the workpiece W or the pallet P is likely to increase at locations farther from the conveyance path 5.

[0064] For example, when the sensor node 10 is attached to a workpiece W that is placed directly on the conveying path 5, the sensor node 10 may be placed above the vertical center of the workpiece W, as shown in Fig. 17. Furthermore, although not shown, when the sensor node 10 is attached to a workpiece W that is placed on the conveying path 5 via a pallet P, the sensor node 10 may be attached above the vertical center of the pallet P.

[0065] This allows the distance between the part of the workpiece W or pallet P where the sensor node 10 is attached and the conveying path 5 to be greater than when the sensor node 10 is not attached in this manner, making it easier for the sensor node 10 to detect vibrations in the conveying path 5.

[0066] Furthermore, when the workpiece W and the sensor node 10 are transported along the transport path 5, for example, in order to process the workpiece W on the transport path 5, the transport of the workpiece W may be stopped or started by a stopper (not shown) provided along the transport path 5.

[0067] In this case, the stopping and starting of the transport of the workpiece W may be controlled by the opening and closing operation of the stopper while the operation of the transport path 5 is maintained. Then, the transport of the workpiece W is suddenly stopped by closing the stopper, and suddenly started by opening the stopper. When the transport of the workpiece W is suddenly stopped or suddenly started by such a stopper, the rear side of the workpiece W in the traveling direction may be lifted up by inertia when the workpiece W suddenly stops or suddenly starts. When the sensor node 10 attached to the workpiece W detects vibrations, this action of lifting the workpiece W by inertia may become vibration noise in the transport path 5 detected by the vibration sensor.

[0068] On the other hand, when the sensor node 10 is transported as a transport body together with the workpiece W, the sensor node 10 may be arranged on the front side of the workpiece W in the traveling direction, assuming that the direction in which the workpiece W is transported by the transport path 5 is the traveling direction. For example, in FIGS. 14 to 17, when the workpiece W is transported from the left side to the right side of the paper, the sensor node 10 may be arranged on the front side of the workpiece W in the traveling direction or on the front surface of the pallet P in the traveling direction, as shown in FIGS. 14 to 17.

[0069] According to this, even when the rear side of the workpiece W in the traveling direction is lifted, the vibration of the sensor node 10 is suppressed, so that the sensor node 10 is less susceptible to the influence of noise caused by the vibration of the workpiece W.

[0070] Alternatively, when it is desired to detect the behavior of the workpiece W whose transportation suddenly stops or starts due to a stopper, the sensor node 10 may be arranged on the rear side of the traveling direction of the workpiece W. For example, in FIG. 18, when the workpiece W is transported from the left side to the right side of the paper, the sensor node 10 may be arranged on the rear side of the traveling direction of the workpiece W, as shown in FIG.

[0071] This structure makes it easy for the sensor node 10 to be lifted together with the workpiece W when the rear side of the workpiece W in the direction of travel is lifted, making it easier to detect the behavior of the sensor node 10 caused by vibrations of the workpiece W.

[0072] The sensor node 10 shown in FIG. 6 is a composite sensor equipped with multiple sensors 11, but the composite sensor is configured with multiple wireless sensor boards 15. Specifically, the composite sensor is configured by arranging wireless sensor boards 15 on one or more surfaces of a hexahedron. As shown in FIG. 7, the wireless sensor board 15 is mounted with one type of sensor 11, a communication unit 13, and other electronic components 15a such as resistors, capacitors, and a microcomputer. Based on power supply from the power supply unit 12, the wireless sensor board 15 has the function of transmitting sensor data indicating the detection results of the sensor 11 to the corresponding receiving unit 20 via the communication unit 13. Note that the number of wireless sensor boards 15 shown in FIG. 7 does not match that shown in FIG. 6. This is for the sake of simplicity, and FIG. 7 only shows the wireless sensor boards 15 located on the front side and the back side of the paper of the wireless sensor boards 15 provided in the sensor node 10.

[0073] A power supply unit 12 is disposed in the center of the composite sensor, and is electrically connected to each wireless sensor board 15, thereby supplying power from the power supply unit 12 and operating the composite sensor. In Figures 6 and 7, the power supply unit 12 is hexahedral in shape, with a wireless sensor board 15 disposed on one of the six faces where sensing is desired, and a power supply terminal 12a is exposed on the face on which the wireless sensor board 15 is disposed. A battery connector 15b that comes into contact with the power supply terminal 12a of the power supply unit 12 is provided on the back side of the wireless sensor board 15, and when the wireless sensor board 15 is attached to the power supply unit 12, power can be supplied to the sensor 11, communication unit 13, etc.

[0074] Furthermore, because the power supply unit 12 is polyhedral in shape and the wireless sensor boards 15 are attached to each face of the power supply unit 12, the power supply unit 12 is positioned in the center of the polyhedral shape of the sensor node 10. By positioning the power supply unit 12 in this way, the volume of the area that becomes the power supply unit 12 can be increased when the number of wireless sensor boards 15 is the same, compared to a non-polyhedral shape. This allows the battery capacity of the power supply unit 12 to be increased, and the operating time of the sensor node 10 to be longer. In other words, it is possible to maximize the operating time of the sensor node 10 while minimizing the shape of the sensor node 10.

[0075] Furthermore, it is preferable that the sensor node 10 be capable of being powered wirelessly so that it can be charged even when the composite sensor is housed in the housing 14. When using wired power supply, one surface of the polyhedral shape constituting the sensor node 10 can be made a surface to which a charging connector can be connected.

[0076] Furthermore, by making all the wireless sensor boards 15 the same size, or by making them smaller than one side of the polyhedron formed by the power supply unit 12, it becomes possible to install the wireless sensor boards 15 on all sides. This allows the wireless sensor boards 15 to be installed on the side that is advantageous for the sensing target of the sensor 11, for example, the side that has high sensitivity. For example, if the sensor node 10 is a regular hexahedron, it becomes a composite sensor that can accommodate six wireless sensor boards 15.

[0077] If the sensor 11 is a microphone that detects sound, by mounting a wireless sensor board 15 on the four sides of the regular hexahedron (front, back, left, and right) in the direction of movement of the sensor node 10, it becomes possible to perform beamforming using array signal processing.

[0078] When the sensor 11 is a temperature sensor or humidity sensor that detects the temperature and humidity of the ambient atmosphere, the ambient atmosphere can be detected by arranging the wireless sensor board 15 on a surface other than the bottom of the regular hexahedron. Furthermore, when the sensor node 10 is installed on the transport path, the surfaces other than the bottom are less susceptible to the temperature of the transport path due to heat transfer. For this reason, it is preferable to arrange the wireless sensor board 15 on the temperature sensor on the surfaces other than the bottom. Furthermore, if two temperature sensors are arranged as the sensor 11 and a flow sensor is also arranged, the flow sensor can measure the air volume around the sensor node 10, and the direction of the wind can be determined from the difference in temperatures detected by the two temperature sensors. This allows for downflow management within the facility 2.

[0079] When sensor 11 is a vibration sensor, installing it on the top surface of the polyhedron increases the moment and improves the sensitivity of the composite sensor to tilt. This makes it possible to detect minute tilts of the workpiece or rattles in the transport path, enabling earlier prediction of abnormalities in equipment 2.

[0080] Furthermore, when the sensor 11 is a temperature sensor, humidity sensor, or vibration sensor, in addition to monitoring the status of the equipment 2, the sensor 11 can be attached to a product and used in a traceability system that manages the product's manufacturing history, post-completion transportation status, and other histories. For example, as shown in FIG. 19 , the equipment status monitoring system 1 includes a server 60 that receives various sensor data detected by the sensor node 10. The server 60 is configured to be able to communicate with the sensor node 10. The server 60 is configured with a microcomputer or the like that includes a CPU, ROM, RAM, flash memory, HDD, and the like (not shown). The server 60 realizes various control operations by having the CPU read and execute programs from the ROM or the like. Note that the storage medium, such as the ROM, is a non-transient physical storage medium. The server 60 in this embodiment functions as a storage unit.

[0081] The server 60 stores, for example, information received from the sensor node 10 as sensor data such as temperature, humidity, and vibration of the ambient atmosphere during the manufacture of the product, in association with the time of receipt of the received information. As a result, as shown in Fig. 20, the equipment status monitoring system 1 can be used as a traceability system that keeps track of the history of various pieces of information during manufacture that indicate the ambient atmosphere under which the product was manufactured.

[0082] Furthermore, when the sensor node 10 is attached to a completed product, information regarding the state of the completed product can be stored in the server 60 by linking the information detected by the sensor node 10 as sensor data regarding the completed product with the time of detection of the information.

[0083] For example, the server 60 can store various types of historical information such as the temperature, humidity, and vibration of the ambient atmosphere during the packaging period from when the product is completed until it is packaged, the storage period from packaging until it is loaded onto a transport vehicle, and the transportation period while the product is being transported by the transport vehicle. This makes it possible to determine the environmental atmosphere under which the product was stored and transported, as shown in Fig. 20. The various types of historical information stored in the server 60 are not limited to temperature, humidity, and vibration, and may include information on sound, acceleration, angular velocity, magnetism, light, surrounding images, flow rate, pressure, odor, etc., depending on the configuration of the sensor 11.

[0084] Furthermore, the various types of history information stored in the server 60 may be displayed on the display device 40 described below so that it can be viewed by the workers 3 involved in the production of the product, or it may be displayed on a display device different from the display device 40. Furthermore, the various types of history information stored in the server 60 may be configured so that it can be viewed by the purchaser who purchased the product. In this way, the various types of information detected by the sensor node 10 can also be used as information stored in the traceability system.

[0085] The server 60 may be provided in the status detection unit 30 described below, or may be provided in the equipment status monitoring system 1 separately from the status detection unit 30. Furthermore, time information possessed by the server 60 itself may be used as the information on the reception time linked to the sensor data received from the sensor node 10. If the sensor node 10 itself is capable of acquiring time information, the information on the reception time linked to the sensor data received from the sensor node 10 may be time information transmitted from the sensor node 10 together with the sensor data. Furthermore, the information on the reception time linked to the sensor data received from the sensor node 10 may be information based on the work content of the worker 3 transmitted from a device (e.g., an RF-ID reader) separate from the sensor node 10 that is operated when the worker 3 performs the work.

[0086] When the sensor node 10 is a polyhedron-shaped composite sensor, if it is desired to use more wireless sensor substrates 15 than the number of faces of the polyhedron, this can be achieved by using multiple sensor nodes 10.

[0087] Furthermore, when multiple sensor nodes 10 are used, for example, as shown in FIG. 21, by configuring an array by arranging microphones on each sensor node 10 at different positions on the transport path 5, the distance and position of the source of the detected sound can be measured based on the distance between the microphones. In this case, the distance between each microphone and the number of microphones arranged in the array are arbitrary and can be selected according to the sound to be detected. For example, if the sound of a facility 2 that is farther away is to be detected, it is advisable to increase the distance between the microphones. Also, to more accurately identify the position of the sound source, it is advisable to increase the number of microphones.

[0088] Furthermore, when the sensor node 10 is configured as a composite sensor, communication between the wireless sensor boards 15 is also possible. In this case, the wireless sensor boards 15 can share a trigger for transmitting sensor data by communicating with each other. For example, the trigger for transmitting sensor data may be the detection result of the sensor 11 exceeding a predetermined threshold. In this case, the trigger is acquired by the wireless sensor board 15 located at the tip of the sensor node 10 in the direction of movement and transmitted to the other wireless sensor boards 15 via communication. To acquire a trigger, the wireless sensor board 15 must process the sensing signal of the sensor 11 and perform various calculations, which consumes power. For this reason, if one wireless sensor board 15 acquires a trigger as the main wireless sensor board 15 and shares it as a trigger for the other wireless sensor boards 15, the other wireless sensor boards 15 can acquire the trigger using only the power consumption required for communication, thereby reducing power consumption.

[0089] Furthermore, it is conceivable that the master wireless sensor board 15 will not start up due to some malfunction. For this reason, it is preferable that when another wireless sensor board 15 detects that a malfunction has occurred in the master wireless sensor board 15 based on communication with the master wireless sensor board 15, the other wireless sensor board 15 will switch over to the new master wireless sensor board 15. In this way, even if a malfunction occurs in the master wireless sensor board 15, trigger acquisition can be continued by the other wireless sensor board 15, and facility monitoring can be performed without stopping sensing by the sensors 11 provided in the wireless sensor boards 15 that are not malfunctioning. Note that the wireless sensor boards 15 referred to here can be implemented in multiple wireless sensor boards 15 provided in one sensor node 10, or in multiple wireless sensor boards 15 provided in multiple sensor nodes 10.

[0090] The size of the sensor node 10 is determined according to the constraints of the transport path to be used and the installation location.

[0091] [Another example of the sensor node 10 configuration] In the configuration shown in FIG. 6, the composite sensor is configured by placing one sensor 11 on one wireless sensor board 15 on each face of a hexahedron, but the composite sensor may also be configured with other structures.

[0092] For example, as shown in FIG. 22, multiple sensors 11 may be mounted on a single board 16. In this figure, a power supply unit 12, for example, consisting of a battery, is also installed on the board 16, and multiple sensors 11 are arranged around the power supply unit 12. As shown in FIG. 23, multiple expansion boards 17a, each with a single sensor 11, may be mounted on a main board 17b. In this figure, the power supply unit 12, consisting of a battery, is arranged in an area of ​​the main board 17b other than the area where the expansion board 17a is mounted, and the expansion boards 17a are arranged around the power supply unit 12. As shown in FIG. 24A, multiple boards 18, each with a single sensor 11 mounted thereon, may be combined as shown in FIG. 24B to form a composite sensor. For example, a storage box 19 capable of slidingly storing multiple boards 18 is prepared, and multiple boards 18 with sensors 11 mounted thereon are stored in the storage box 19. In this case, as shown in FIG. 24A, each substrate 18 may be provided with a power supply unit 12 consisting of a battery or the like, or a power supply unit 12 may be provided on some of the multiple substrates 18, and power may be supplied to the other substrates 18 as well.

[0093] However, if there is a size restriction, there will be restrictions on the power supply from the power supply unit 12, the number of sensors 11, etc., so it is preferable to use a polyhedron shape as shown in FIG.

[0094] The wireless sensor board 15 and each board equipped with the sensor 11 shown in the other configuration examples may have a self-diagnosis function for each sensor 11. Having a function to diagnose whether the sensor sensitivity of the sensor data during learning and that during operation are the same, whether the sensor itself is malfunctioning, etc., can increase the reliability of determining the degree of abnormality. Furthermore, based on such a self-diagnosis function, it is also possible to provide, for example, a temperature correction function. Because the sensor 11 has temperature characteristics, it is possible to correct the sensor sensitivity according to the environmental temperature. In other words, even in an environment with temperature changes such as a circulating furnace, temperature correction based on the self-diagnosis function enables more accurate sensing, which is effective.

[0095] [Location determination of sensor node 10] If the sensor node 10 is installed in or near the facility 2 and not on a mobile object, the installation location is identified as the location of the sensor node 10.

[0096] When the sensor node 10 is installed on a moving object, for example, when the sensor node 10 is installed as a moving object on a transport path 5, it is necessary to identify the location of the moving object. To identify the location of the moving object, for example, if the moving object is installed on a transport path 5 that moves at a constant speed, the location can be identified using a time trigger. That is, since the movement distance of the sensor node 10 can be determined based on the movement speed, the location of the sensor node 10 can be identified by measuring the elapsed time from the start of the sensor node 10's movement. Then, for example, a sound sensor can be used as the sensor 11 to measure the direction from which sound is coming, thereby pinpointing the desired detection location. Alternatively, for example, an optical sensor can be used as the sensor 11 to measure the direction from which light is shining and identify the location of the sensor node 10 based on the amount of light received. If the transport path does not move at a constant speed, the location of the sensor node 10 can be identified using an image analyzer, RF-ID reader, optical marker, or the like as the sensor 11.

[0097] Alternatively, a speaker that emits a specific sound, such as a sine wave sound, at a constant sound pressure can be installed on the equipment 2 side, and the sound pressure of that frequency can be maximized when the sensor node 10 is closest to the speaker. For example, as shown in Figure 25, consider a case where the sensor node 10 is installed on a transport path 5. A speaker 6, which acts as a sound source, is installed near the transport path 5. In this case, as indicated by the arrow in the figure, when the sensor node 10 is moved from the left side to the right side of the page on the transport path 5, the sound pressure maximizes near the speaker 6. Specifically, if a 2000 Hz sound is output from the speaker 6, the sound pressure near 2000 Hz is measured as shown in Figure 26. In this figure, the sound pressure reaches its highest point at 7.5 seconds, which is the timing when the sensor node 10 is closest to the speaker 6, and the position of the sensor node 10 can be identified.

[0098] Furthermore, ultrasonic waves can be used to distinguish between sounds that are actually being generated and sounds that are not generated by the speaker 6. In this case, a high-frequency microphone can be installed as the sensor 11.

[0099] Furthermore, when the moving conveyance path 5 is divided into multiple sections as shown in FIG. 27, the position of the conveyance path 5 may be identified based on normal data of the conveyance path 5 learned by the state detection unit 30 described below. For example, assume that the conveyance path 5 is divided into a first conveyance section R1, a second conveyance section R2, a third conveyance section R3, and a fourth conveyance section R4 as shown in FIG. 27. In this case, the state detection unit 30 learns various data representing the state of the conveyance path 5 in each section from the first conveyance section R1 to the fourth conveyance section R4 and stores a model. The position of the conveyance path 5 may then be identified by comparing the model with sensor data transmitted from the sensor node 10. The various data representing the state of the conveyance path 5 may include vibration, acceleration, angular velocity, temperature, humidity, electromagnetic field, sound, light intensity, force, torque, and surrounding images detected by the sensor node 10.

[0100] In this way, when the state detection unit 30 determines the position of the transport path 5 based on the data of the transport path 5 learned, it is possible to eliminate the need for a device that generates a sound source or light source for the sensor node 10 to determine the position.

[0101] [Receiving unit 20] The receiving unit 20 receives sensor data transmitted from the sensor node 10, and various signals transmitted from the facility 2, such as facility memory signals and facility operation signals. In Fig. 1, the receiving unit 20 and the state detecting unit 30, which will be described later, are shown as separate components, but the receiving unit 20 and the state detecting unit 30 can also be configured by a device that has both a receiving function and various arithmetic processing functions, such as a personal computer.

[0102] [Status detection unit 30] The status detection unit 30 detects the status of each component of the equipment 2 to be monitored, detects whether there is an abnormality or a sign of an abnormality for each component of the equipment 2, and outputs the results to a display device 40 or the like. For example, the status detection unit 30 learns data on each component when each piece of equipment 2 is operating normally and stores a model, and when monitoring for an abnormality, it acquires data on each component when the equipment 2 is in use and compares it with the learned model to detect the status of each component. The status detection unit 30 has such a function for each component to be detected.

[0103] FIG. 28 is a block diagram showing details of the functional blocks of the state detection unit 30.

[0104] As shown in this figure, the status detection unit 30 is configured to have multiple machine learning units 31, one for each component to be detected, and a signal output unit 32. While Fig. 28 shows the details of the functional blocks of only one of the multiple machine learning units 31, in reality, multiple units with similar block configurations are provided. The machine learning units 31 make inferences about the occurrence of an abnormality or its signs in the equipment 2, and the inference results from the machine learning units 31 for each component are collectively processed by a single signal output unit 32, thereby monitoring an abnormality in each piece of equipment 2.

[0105] The components to be detected here are elements that can be a cause of abnormalities in the equipment 2 to be monitored, and are elements to be detected based on sensor data. The components may be specific locations in the equipment 2 determined by the worker 3, or sections separated by each piece of equipment 2. When determining the components to be detected by the worker 3, relying on locations that are checked by the intuition and knack of an expert makes it easier to identify areas that are prone to breakage. Also, having an expert wear glasses that can detect line of sight and observing the inspection work, etc., is useful for visualizing the intuition and knack.

[0106] The machine learning unit 31 includes a state observation unit 31a, a label data estimation unit 31b, a learning unit 31c, a model storage unit 31d, and an inference result output unit 31e.

[0107] The state observation unit 31a inputs sensor data transmitted from the sensor node 10, observes it as a state variable representing the state of the component to be detected, and transmits the observation data to the learning unit 31c and the inference result output unit 31e. The state observation unit 31a can also input detection results indicated by sensing signals from various sensors 2a originally installed in the facility 2 as sensor data. In this case, the state observation unit 31a treats the detection results of the installed sensors 2a in the same way as sensor data transmitted from the sensor node 10 and observes them as a state variable representing the state of the component to be detected. Examples of physical quantities and states detected by the installed sensors 2a include voltage, current, positional displacement, speed, vibration or acceleration, temperature, humidity, electromagnetic field, sound, light intensity, force, torque, surrounding images, distance, flow rate, pH, pressure, viscosity, and odor. The sensors 2a installed in the facility 2 may be either composite sensors or single sensors, and communication with the state detection unit 30 may be wired or wireless.

[0108] The label data estimation unit 31b acquires an equipment memory signal and an equipment operation signal, which are actual operating state data of the equipment 2, as label data and transmits them to the learning unit 31c and the like. The equipment memory signal is a signal indicating how the equipment 2 was processed. When an abnormality is detected in the equipment 2 and the abnormality is handled according to the detection result of the equipment status monitoring system 1, the label data estimation unit 31b stores the signal as an equipment memory signal. In addition, when the operator 3 handles the abnormality directly based on intuition or skill without following the detection result, the label data estimation unit 31b also stores the signal as an equipment memory signal. Then, in order to feed back this history, the equipment memory signal is transmitted to the label data estimation unit 31b. The equipment operation signal is a signal indicating what operation the equipment 2 performed when an abnormality handling process was performed on the equipment 2. Since the equipment operation signal indicates what state the equipment 2 entered when it was handled, the signal is label data associated with the equipment memory signal and labeled.

[0109] For example, the label data inference unit 31b acquires, as operation status data, an operation trigger of the equipment 2, which can be acquired by a PLC (programmable logic controller) or the like. Alternatively, the 5M1E (person, equipment, material, method, measurement, and environment) of the worker 3 may be acquired as operation status data. For example, when an abnormality occurs in the equipment 2 on a certain date, the equipment memory signal indicates which part of the equipment 2 is abnormal, how, and why, as well as which worker 3 responded and how. The equipment operation signal indicates how the equipment 2 behaves as a result. The label data acquired by the label data inference unit 31b may be used only for learning by the learning unit 31c (described later), or may also be used during inference by the inference result output unit 31e. When used for inference by the inference result output unit 31e, the label data is transmitted to the inference result output unit 31e.

[0110] The learning unit 31c creates a model that estimates the degree of anomaly of the component to be detected based on the state variables indicated by the observation data from the state observing unit 31a and the operating state of the equipment 2 indicated by the label data from the label data estimating unit 31b. For example, the learning unit 31c creates a model for normal times based on various physical quantities and the operating state of the equipment 2 when the equipment 2 to be monitored is operating normally. The learning unit 31c may also create a model for abnormal times based on various physical quantities and the operating state of the equipment 2 when an abnormality occurs.

[0111] The learning data used by the learning unit 31c may include characteristic portions extracted based on at least one of the fluctuation amount, amplitude, fluctuation time, number of fluctuations, frequency, and deviation from a specified value that outputs a signal indicating an abnormality, as well as time-series data. Only one piece of sensor data may be used for learning, or a set of characteristic portions of multiple physical quantities or time-series data may be used as a state variable to be learned and estimated. This also includes time-series transitions of characteristic quantities obtained by machine learning. For example, this includes characteristic quantities obtained by dimensionality reduction using unsupervised machine learning such as principal component analysis or t-SNE (T-distributed Stochastic Neighbor Embedding). Furthermore, the learning unit 31c may use multiple physical quantities and weight them based on past data for learning. Furthermore, by learning the operating status of the equipment 2 acquired using label data, a model that is more specifically targeted to the locations and operations that require monitoring can be created. Even if label data is not available, only the observation data from the state observation unit 31a may be used as learning data.

[0112] The model storage unit 31d is a unit that stores the model created by the learning unit 31c, i.e., the learning data that serves as the model, as a reference model. Specifically, the learning unit 31c stores a model for when the equipment 2 to be monitored is normal as a reference model for estimating the degree of abnormality of the component to be detected. Furthermore, if the learning unit 31c has also created a model for when an abnormality occurs, this model is also stored in the model storage unit 31d as a reference model for when an abnormality occurs.

[0113] The inference result output unit 31e infers the operating state of the equipment 2 to be monitored at the time of monitoring, based on the learning data of the stored model. When inferring the operating state of the equipment 2, the inference result output unit 31e can input observation data and label data, and infer the operating state of the equipment 2 using the observation data and label data in addition to the learning data of the stored model. The operating state here means the degree of deviation from the normal state, that is, how far it deviates from the learning data for normal times. The inference result output unit 31e quantifies the degree of deviation as an "abnormality degree" and outputs it as the inference result.

[0114] For example, the "degree of abnormality" value is a judgment value obtained by statistically processing the change in the raw value or the value of the physical quantity of the sensor data sent from the sensor node 10 during monitoring. The "degree of abnormality" may be a judgment value or a change in the raw value obtained from one physical quantity detected from one sensor 11, or may be a composite judgment value or a change in the raw value based on multiple physical quantities detected from multiple sensors 11.

[0115] The "degree of abnormality" can be inferred not only as a current value, i.e., a value used to determine whether an abnormality is currently occurring in the equipment 2, but also as a value expected in the future, i.e., a value used to predict an abnormality in the equipment 2. The current "degree of abnormality" is calculated, for example, by comparing current observation data with learning data. The expected "degree of abnormality" can be calculated by estimating future observation data from current observation data and comparing the expected observation data with learning data, and can also be calculated from the current "degree of abnormality." The expected "degree of abnormality" can be calculated according to elapsed time by learning the association between past operating state data and the state of the equipment 2 in the model storage unit 31d. The inference result output unit 31e then outputs the inference result to the signal output unit 32.

[0116] The signal output unit 32 determines whether an abnormality has occurred in the equipment 2 or whether there is a sign of an abnormality based on the "degree of abnormality" transmitted from the inference result output unit 31e, and transmits the determination result to the display device 40. For example, the signal output unit 32 pre-stores a threshold value corresponding to the "degree of abnormality" indicated by a determination value obtained by performing statistical processing on changes in raw values ​​of sensor data or values ​​of physical quantities. The signal output unit 32 then determines that an abnormality has occurred or a sign of an abnormality has occurred when the value of the "degree of abnormality" exceeds the pre-stored threshold value. The sign of an abnormality can be used to infer not only the possibility that an abnormality may occur in the future, but also the remaining time until the abnormality occurs. For example, as described above, by calculating the "degree of abnormality" according to elapsed time, the elapsed time required for the "degree of abnormality" to exceed the threshold can be determined. Based on this, the remaining time until the abnormality occurs can be inferred.

[0117] The signal output unit 32 can also infer the location of the abnormality, that is, the component in which the abnormality occurred. From the "degree of abnormality" described above, it is possible to identify which component of the equipment 2 has experienced an abnormality or a sign of an abnormality. Furthermore, by looking at the value of the degree of abnormality for each component, the signal output unit 32 can determine that the component with the highest degree of abnormality is the fault location, and can identify the location where the sign of an abnormality has occurred.

[0118] For example, if a sound sensor is used as the sensor 11 and an abnormality feature is detected in the sound sensor data, focusing on that feature allows the direction of the sound source to be identified using multiple microphones, enabling more precise identification of the location of the abnormality sign. For example, as shown in FIG. 29, consider a case in which the sensor node 10 is moved along the transport path 5 and passes in front of each piece of equipment 2, with the equipment 2 at position (3) emitting an abnormal sound. In this case, as the sensor node 10 moves from position (1) to position (6), the sound sensor detects the abnormal sound. However, the abnormal sound is detected as a small sound at positions farther from position (3), and as the position gets closer to position (3), the sound is detected as a louder sound. Therefore, the relationship between the position indicated by the sound sensor detection results and the "degree of abnormality" is shown as shown in FIG. 30, which indicates that the abnormal sound is being emitted from the equipment 2 at position (3), which has the highest "degree of abnormality." The inference result output unit 31e calculates the "degree of abnormality" at each position based on the abnormality feature that appears in the sound sensor data—in the example of FIG. 29, the loudness of the sound. Therefore, by comparing the "degree of abnormality" calculated by the inference result output unit 31e with a threshold value assumed when an abnormal sound is emitted, it is possible to determine that an abnormality has occurred in the equipment 2 at the position (3) where the "degree of abnormality" exceeds the threshold value. Note that although a sound sensor is used as an example here, even if a sensor other than a sound sensor is used, it is possible to similarly focus on the feature amount obtained from sensor data and, when a feature amount of an abnormality is detected, identify the location of the abnormality.

[0119] [Display device 40] The display device 40 displays information according to the determination result transmitted from the signal output unit 32, and is configured, for example, with a display. When the signal output unit 32 notifies the display device 40 that an abnormality has occurred or that there is a sign of an abnormality, the display device 40 displays this information. Furthermore, when the signal output unit 32 transmits the determination result that there is no abnormality in the equipment 2, the display device 40 can also display this information.

[0120] The display method used by the display device 40 can be set arbitrarily by the worker 3. The location of the abnormality or the location of the potential abnormality may be displayed by name, but a 3D mapping display allows the worker 3 to intuitively grasp the relevant location. Furthermore, if a display using AR (Augmented Reality) is performed, the worker 3 can confirm the location of the potential abnormality or visually confirm the recommended repair while maintaining the equipment 2. Furthermore, if a sound sensor is used in the sensor node 10, the display device 40 may be configured to output the sound of the equipment 2 detected by the sound sensor. This allows the worker 3 to check for abnormal sounds and audibly confirm the abnormality of the equipment 2 while maintaining the equipment 2.

[0121] For example, as shown in FIG. 31 , the left side of the display constituting the display device 40 displays the components 2b to be detected in the facility 2 in a form that allows the operator 3 to grasp the components 2b. The components 2b in the facility 2 shown here may be locations that are to be detected in each of multiple different pieces of facility 2, or may be multiple locations that are to be detected in the same facility 2. The right side of the display shows the "degree of abnormality" corresponding to each component 2b and the corresponding elapsed time. In this way, the display device 40 allows the operator 3 to grasp how the "degree of abnormality" of the components 2b in the facility 2 to be monitored will change in the future.

[0122] 32 and 33, details of another example of the content displayed by the display device 40 when the equipment status monitoring system 1 detects the occurrence of an abnormality or a sign of an abnormality in the equipment 2 will be described. In the example shown in these figures, a triaxial angular velocity sensor is applied to the sensor 11 so that the equipment status monitoring system 1 can detect the occurrence of an abnormality or a sign of an abnormality in each of the first conveying path 51, the second conveying path 52, the third conveying path 53, and the fourth conveying path 54 that are successively connected. The sensor node 10 detects angular velocities in three directions that are orthogonal to each other on each of the first conveying path 51 to the fourth conveying path 54 when the object is conveyed in this order on the first conveying path 51, the second conveying path 52, the third conveying path 53, and the fourth conveying path 54.

[0123] The machine learning unit 31 creates a model for estimating the degree of abnormality for each of the first to fourth transport paths 51 to 54 from information relating to the operating status of each of the first to fourth transport paths 51 to 54 detected by the sensor node 10. Then, the display device 40 displays the occurrence of an abnormality or the occurrence of a sign of an abnormality for each of the first to fourth transport paths 51 to 54 as a detection result.

[0124] Specifically, the display device 40 displays detection results corresponding to the time taken for the sensor node 10 to travel on each of the transport paths 51, 52, 53, and 54 when it is transported from the first transport path 51 to the fourth transport path 54. The displayed content of the detection results may be information on the change over time in the angular velocity in each of the three directions detected by the triaxial angular velocity sensor, or may be a three-dimensional model showing the change over time in the attitude of the sensor node 10 calculated based on the angular velocities in each of the three directions detected by the triaxial angular velocity sensor. When the displayed content is a three-dimensional model showing the change over time in the attitude of the sensor node 10, the worker 3 can more visually grasp the abnormal state of the equipment 2 more easily.

[0125] Incidentally, when calculating the change over time in the attitude of the sensor node 10 into a three-dimensional model, there is a risk of an error occurring between the actual attitude of the sensor node 10 and the attitude of the sensor node 10 calculated using the three-dimensional model. The longer the detection period of the sensor node 10 used to calculate the three-dimensional model, the more likely this error will accumulate and become larger.

[0126] For this reason, an angular velocity sensor different from the triaxial angular velocity sensor that the sensor node 10 has may be added to the sensor node 10 to correct the error, or a sensor different from the angular velocity sensor (for example, a triaxial geomagnetic sensor) may be added to the sensor node 10 to correct the error. Furthermore, if there is a point where the posture of the sensor node 10 is uniquely determined when the sensor node 10 is transported in this order from the first transport path 51 to the fourth transport path 54, a three-dimensional model of the sensor node 10 may be calculated based on this uniquely determined posture of the sensor node 10 to reduce the accumulated error.

[0127] Furthermore, the display device 40 may display the detection results of the first to fourth transport paths 51 to 54, and may also display an image of the sensor node 10 captured by the image sensor, as shown in FIG.

[0128] In addition, if it is possible to detect the occurrence of an abnormality or a sign of an abnormality in each of the first conveying path 51, the second conveying path 52, the third conveying path 53, and the fourth conveying path 54, the angular velocity sensor applied to the sensor 11 may be a two-axis angular velocity sensor or a one-axis angular velocity sensor.

[0129] The display device 40 may also be capable of displaying the change in sensor data over time when an abnormality or a sign of an abnormality is detected. For example, when a sign of an abnormality is detected in the first transport path 51, the display device 40 may display the sensor data of the first transport path 51 detected by the sensor node 10 for a predetermined period from the time of detection to a time going back, for example, one hour, one day, or one month. In this case, the display device 40 may be configured so that the operator 3 can play and stop the display of the change in sensor data over time by operating a playback start switch 41 and an operation stop switch 42 displayed on the display. The display device 40 may also be configured so that the operator 3 can adjust the magnification of the playback speed when displaying the change in sensor data over time by operating a playback speed adjustment switch 43 displayed on the display.

[0130] In this way, by displaying the sensor data detected by the sensor node 10 on the display device 40, the worker 3 can visually confirm changes in the posture of the sensor node 10. However, when the equipment status monitoring system 1 of this embodiment detects the occurrence of an abnormality or a sign of an abnormality in the equipment 2, it is considered that the equipment 2 needs to be promptly inspected by the worker 3. When the worker 3 inspects the equipment 2, it may be necessary to stop the operation of the equipment 2.

[0131] However, even if inspection of equipment 2 is not necessarily required, it is conceivable that equipment status monitoring system 1 may detect an abnormality or a sign of an abnormality occurring in equipment 2 due to a change in the external environment, etc. For this reason, even if equipment status monitoring system 1 detects an abnormality or a sign of an abnormality occurring in equipment 2, it is desirable not to stop operation of equipment 2 if inspection of equipment 2 is not required.

[0132] On the other hand, even if the equipment status monitoring system 1 detects the occurrence of an abnormality or a sign of an abnormality in the equipment 2, the worker 3 can check the sensor data displayed on the display device 40 and determine whether it is necessary to inspect the equipment 2. For example, when the equipment status monitoring system 1 detects a sign of an abnormality in the equipment 2, the worker 3 can easily determine whether it is necessary to inspect the equipment 2 by checking the change in the sensor data over time. This makes it possible to avoid unnecessary shutdowns of the equipment 2 and reduce equipment downtime, thereby improving production efficiency.

[0133] Furthermore, because abnormalities in the equipment 2 do not occur on a daily basis, the equipment status monitoring system 1 is unlikely to detect the occurrence of an abnormality or the occurrence of a sign of an abnormality on a daily basis. However, by having the worker 3 visually or audibly check the sensor data detected by the sensor node 10 on the display device 40 on a daily basis, even if the worker 3 who performs the inspection is a beginner, it becomes easier for the worker 3 to determine the need to inspect the equipment 2. In other words, the equipment status monitoring system 1 can be used to train the workers 3 who inspect the equipment 2.

[0134] In particular, the judgment of abnormalities in the equipment 2 is influenced by the intuitive judgment of the worker 3, and is therefore easily influenced by the level of proficiency of the worker 3. However, by having the worker 3 visually or audibly confirm the sensor data used to judge abnormalities in the equipment 2, it becomes easier to convey the intuitive judgment of the highly skilled worker 3 to the less skilled worker 3. In other words, the equipment status monitoring system 1 can educate the less skilled worker 3 about the intuition and tips that the highly skilled worker 3 has intuitively grasped.

[0135] [Operation of equipment status monitoring system 1] As described above, the facility status monitoring system 1 is configured. Next, the operation of the facility status monitoring system 1 configured in this manner will be described.

[0136] First, when the equipment 2 to be monitored is operating normally, the sensor node 10 transmits sensor data consisting of sensing signals from the sensor 11 and the like. This is received by the receiving unit 20 and transmitted to the status detection unit 30. If the equipment 2 is equipped with a sensor 2a, the detection results indicated by the sensing signal of that sensor 2a can also be input as sensor data. On the other hand, when the equipment 2 is to be operated, the worker 3 will drive the equipment 2, and an equipment memory signal and an equipment operation signal, which serve as operating status data at that time, are input to the status detection unit 30.

[0137] As a result, the sensor data is input to the state observation unit 31a, and the equipment memory signal and the equipment operation signal are input to the label data estimation unit 31b. These are then transmitted to the learning unit 31c, whereby the data of each component when the equipment 2 is operating normally is learned, a model is created, and the model is stored in the model storage unit 31d. Since the state of the equipment 2 can also be learned from the label data, it is possible to create a model that is more limited to the locations and operations that are more desired to monitor. Furthermore, if the sensor node 10 is installed on a mobile body such as the transport path 5 and can be moved, the position of the sensor node 10 is also identified, and a model is created that corresponds to the position of the sensor node 10 when the sensor data was obtained.

[0138] Once the model is stored in the model storage unit 31d, the sensor node 10 is used to monitor the occurrence of an abnormality or a sign of an abnormality in the equipment 2 to be monitored. That is, sensor data from the sensor node 10, and further sensor data from sensors 2a attached to the equipment 2 as needed, are transmitted to the state detection unit 30, and data of each component 2b indicated by the sensor data is transmitted to the inference result output unit 31e. As a result, the inference result output unit 31e compares the data of each component 2b with the learning data that serves as the model, calculates the "degree of abnormality" of each component 2b, and also calculates the "degree of abnormality" according to the elapsed time in the future, and transmits it to the signal output unit 32.

[0139] Next, the signal output unit 32 compares the "degree of abnormality" of each component 2b transmitted from the inference result output unit 31e with a corresponding threshold value stored in advance. If the current "degree of abnormality" exceeds the threshold value, it is determined that an abnormality has occurred in the equipment 2, and if the future "degree of abnormality" exceeds the threshold value, it is determined that a sign of an abnormality has occurred, indicating when an abnormality may occur in the equipment 2.

[0140] In this way, when the signal output unit 32 makes a determination, the determination result, such as whether an abnormality has occurred or a sign of an abnormality has been detected, is transmitted to the display device 40, and is displayed by the display device 40. If no abnormality has occurred, a display is made indicating that each piece of equipment 2 is normal. Furthermore, if an abnormality has occurred or a sign of an abnormality has occurred, the equipment 2 is displayed, and the location of the abnormality or the location of the abnormality sign is displayed using a 3D mapping display or the like. Furthermore, if a sign of an abnormality has occurred, the display device 40 also displays the time remaining until the abnormality occurs.

[0141] This allows the worker 3 to check whether the equipment 2 is normal or abnormal based on the display on the display device 40, and if an abnormality has occurred or there are signs of an abnormality, it becomes possible to take action.

[0142] Furthermore, since it is possible to identify the location of an abnormality or the location of a possible abnormality, it is possible to estimate the parts that need to be replaced and automatically order the replacement parts from the manufacturer. Regarding the signs of an abnormality, since it is possible to know when an abnormality may occur, it is also possible to set a delivery date for the replacement parts according to the time of the abnormality occurrence and order the parts. In this way, it is possible to avoid holding unnecessary inventory and to prepare for maintenance before an abnormality occurs.

[0143] For example, as shown in Fig. 34, the status detection unit 30 sets a delivery date for parts manufacturer A that manufactures the replacement parts, and places an order for the replacement parts. This allows parts manufacturer A to determine delivery dates and place orders with parts manufacturers B and C that manufacture parts necessary for manufacturing the replacement parts, so that the replacement parts can be delivered in time for the delivery date. Then, parts manufacturers B and C can also place orders with related parts manufacturers so that they can deliver their parts in time for parts manufacturer A's delivery date. In this way, parts can be ordered in advance from each parts manufacturer related to the replacement parts.

[0144] Furthermore, if the same product is manufactured on multiple lines, the target production volume for each line can be reset based on the signs of anomalies. For example, if signs of anomalies are detected on one line, the target production volume for each line can be reset so that the target daily or monthly production volume can be achieved with the minimum operating time, taking into account the equipment downtime for maintenance. This makes it possible to set an appropriate target production volume that takes into account the signs of anomalies.

[0145] Furthermore, if a correlation diagram between total energy consumption and production volume is acquired in a factory where the equipment status monitoring system 1 is applied, it is possible to isolate the causes of energy usage based on the relationship between production volume and energy consumption. For example, as shown by the line graph in Figure 35, suppose there is a situation in which the number of production units increases or decreases due to the operation and shutdown of some equipment 2, but energy usage does not change, as shown by the bar graph. Specifically, when equipment 2 is operating, the number of production units and energy usage are correlated, as in states (1) and (3). When equipment 2 is shut down, energy usage decreases in line with the decrease in production volume, as in state (4). However, if energy usage does not decrease despite the decrease in production volume due to the shutdown of equipment 2, as in state (2), it is possible that the standby power consumption of equipment 2 that is not involved in the stopped production is high.

[0146] In such a case, the status detection unit 30 will show a large value for the "degree of abnormality" of the component 2b provided in the relevant equipment 2, and the signal output unit 32 will determine that an abnormality has occurred. However, the occurrence of an abnormality can also be detected from the relationship between the number of units produced and the amount of energy used.

[0147] Furthermore, when feedback control is used to maintain a constant operation of equipment 2, there is a correlation between the number of units produced and energy consumption, but even so, energy consumption may increase more than expected. For example, there may be cases where the output of equipment 2 gradually increases due to disturbances such as high friction caused by insufficient lubrication or contamination.

[0148] Therefore, if the equipment status monitoring system 1 detects the time-series changes in the number of production units and energy consumption of the equipment 2, it can lead to identifying the cause of the time-series changes in the energy consumption.

[0149] As described above, in the equipment status monitoring system 1 of this embodiment, at least one common sensor node 10 is used to transmit normal sensor data indicating that the plurality of pieces of equipment 2 are operating normally to the status detection unit 30. Then, the status detection unit 30 is made to learn the normal states of the plurality of pieces of equipment 2 as learning data. Therefore, by comparing the states of the plurality of pieces of equipment 2 indicated by the sensor data transmitted from the sensor node 10 after learning with the learning data, it is possible to detect the occurrence of an abnormality or a sign of an abnormality in the plurality of pieces of equipment 2 that are the monitoring targets, without providing a vibration sensor or the like in each of the monitoring targets.

[0150] Furthermore, the state of the component 2b of the equipment 2 is expressed as an "abnormality degree," and it is detected that each component 2b is abnormal or has signs of abnormality. This makes it possible to identify which component 2b of which equipment 2 among multiple pieces of equipment 2 has an abnormality or is showing signs of an abnormality.

[0151] Furthermore, when there is a transport path 5 for transporting products, such as in a production facility, the sensor node 10 is used as a transport body that moves along with the transport path 5 so that the number of facilities 2 to be monitored can be increased. This makes it possible to monitor the state of the facilities 2 from the start to the end of the production of products in the production facility. It is then possible to monitor multiple facilities 2 provided as such a production facility using at least one common sensor node 10.

[0152] Furthermore, by using a composite sensor as the sensor 11 provided in the sensor node 10 and performing composite processing using multiple sensor data, it becomes possible to more accurately detect the occurrence of an abnormality or its precursor in the equipment 2. An example of composite processing is performing processing that includes correlation between each sensor data.

[0153] The Paris Agreement, an international initiative on climate change that began implementation in 2020, calls for efforts to reduce emissions in order to achieve carbon neutrality in the second half of this century. As part of this effort, there has been a growing movement to reduce CO2 emissions from factories during the product manufacturing process to zero. Eliminating wasteful production, such as equipment downtime due to sudden breakdowns and maintenance, is important for reducing CO2 emissions. This can be achieved by detecting abnormalities and their signs in equipment 2 using the equipment status monitoring system 1 described in this embodiment. Specifically, application of the equipment status monitoring system 1 enables parts to be ordered and supplied before an abnormality occurs, eliminating unnecessary inventory and minimizing equipment downtime due to maintenance, thereby significantly contributing to reducing CO2 emissions.

[0154] By transporting the composite sensor on the transport path 5, it is possible to monitor the status of multiple pieces of equipment 2 with just one composite sensor, without having to install a sensor in each piece of equipment 2. Furthermore, the combination of types of sensors 11 that make up the composite sensor can be selected as appropriate, and by changing the placement location of the sensors 11 depending on the type, it is possible to maximize the sensor performance. In particular, if the composite sensor is configured with multiple wireless sensor boards 15, it is possible to easily realize a structure that maximizes sensor performance based on the placement location and combination of the wireless sensor boards 15.

[0155] 6 and 7, that is, the composite sensor is made up of multiple wireless sensor boards 15 and is shaped like a polyhedron, with a wireless sensor board 15 disposed on at least one of its faces. The sensor node 10 has a wireless sensor board 15 mounted with one type of sensor 11 and a communication unit 13, and has the function of transmitting sensor data to the corresponding receiving unit 20 when power is supplied from the power supply unit 12. The power supply unit 12 is disposed in the center of the composite sensor, and is connected to the wireless sensor board 15 to supply power, thereby enabling the composite sensor to operate.

[0156] Such a configuration allows the battery capacity of the power supply unit 12 to be increased, thereby extending the operating time of the sensor node 10. In other words, it is possible to maximize the operating time of the sensor node 10 while minimizing the size of the sensor node 10.

[0157] (Other embodiments) Although the present disclosure has been described based on the above-described embodiment, it is not limited to the embodiment and encompasses various modifications and modifications within the equivalent range. In addition, various combinations and forms, as well as other combinations and forms including only one element, more than one element, or less than one element, are also within the scope and spirit of the present disclosure.

[0158] For example, although the configurations of the sensor node 10 shown in Figures 6, 22, 23, and 24B have been described as examples, the sensor node 10 may have a different configuration from those shown in these figures. For example, while Figure 6 shows an example in which the sensor node 10 is configured as a regular hexahedron, the sensor node 10 may have another polyhedral shape. Also, although Figure 6 and other figures show a structure in which both the sensor 11 and the communication unit 13 are mounted on each wireless sensor board 15, the wireless sensor boards 15 do not need to have the same structure. For example, a structure in which only one communication unit 13 is provided for multiple wireless sensor boards 15 and sensor data from multiple sensors 11 is transmitted by a single communication unit 13 may also be used.

[0159] Furthermore, in the above embodiment, an example of an equipment status monitoring system 1 that monitors multiple pieces of equipment 2 has been given. However, it is acceptable for there to be multiple monitored objects, and different components within a single piece of equipment 2 may be monitored. For example, different parts of the same piece of equipment 2, such as an XY stage and a machining head, may be monitored as monitored components. It is also possible to monitor the status of other systems using the trained model and inference results of the equipment status monitoring system 1 described here. For example, when applying the equipment status monitoring system 1 to the same monitored objects, learning can be performed using the monitored objects for a model used to build the system, and this can be used to monitor the status of other systems.

[0160] Furthermore, the equipment status monitoring system 1 described in the above embodiment does not necessarily require that each component be provided in one location. For example, the sensor node 10, receiving unit 20, and status detection unit 30 are provided within a factory that includes the equipment 2, and the display device 40 is provided outside the factory. Data indicating the results from the status detection unit 30 can then be transmitted to an external cloud or the like, and the data can be imported from the cloud into the display device 40. Even in this configuration, the equipment status monitoring system 1 can still be established. [Explanation of symbols]

[0161] 1...equipment status monitoring system, 2...equipment, 2a...sensor, 2b...component, 3...worker 3 5...Transport path, 10...Sensor node, 12...Power supply unit, 13...Communication unit, 14...Case 15...wireless sensor board, 20...receiving unit, 30...state detection unit, 31...machine learning unit 31a...state observation unit, 31b...label data estimation unit, 31c...learning unit 31d...model storage unit, 31e...inference result output unit, 32...signal output unit 40…Display device

Claims

1. An equipment status monitoring system, comprising: a sensor node (10) that has a sensor (11) that outputs data indicating the state of a facility (2) to be monitored as sensor data, a communication unit (13) that transmits the sensor data, and a power supply unit (12) that supplies power to the sensor and the communication unit, and that is commonly applied to a plurality of the monitored facilities; a receiver (20) that receives the sensor data transmitted from the communication unit; a state detection unit (30) that inputs the sensor data received by the receiver, learns the normal states of the plurality of monitored objects as learning data based on the sensor data in normal times when the plurality of monitored objects are operating normally, and when the sensor data transmitted from the sensor node after learning is received by the receiver, compares the states of the plurality of monitored objects indicated by the sensor data with the learning data to detect the occurrence of an abnormality or a sign of an abnormality in the plurality of monitored objects, a power supply method from the power supply unit to the sensor and the communication unit is one of a method using a magnetic field, energy harvesting, and mobile battery power supply; the sensor node is disposed on a mobile body and moved together with the mobile body to acquire the sensor data indicating the states of the plurality of monitored objects; the moving body is a chuck device that clamps at least one of a workpiece and the sensor node, The sensor detects operation of the chuck equipment.

2. An equipment status monitoring system, comprising: a sensor node (10) that has a sensor (11) that outputs data indicating the state of a facility (2) to be monitored as sensor data, a communication unit (13) that transmits the sensor data, and a power supply unit (12) that supplies power to the sensor and the communication unit, and that is commonly applied to a plurality of the monitored facilities; a receiver (20) that receives the sensor data transmitted from the communication unit; a state detection unit (30) that inputs the sensor data received by the receiver, learns the normal states of the plurality of monitored objects as learning data based on the sensor data in normal times when the plurality of monitored objects are operating normally, and when the sensor data transmitted from the sensor node after learning is received by the receiver, compares the states of the plurality of monitored objects indicated by the sensor data with the learning data to detect the occurrence of an abnormality or a sign of an abnormality in the plurality of monitored objects, a power supply method from the power supply unit to the sensor and the communication unit is one of a method using a magnetic field, energy harvesting, and mobile battery power supply; the sensor node is disposed on a mobile body and moved together with the mobile body to acquire the sensor data indicating the states of the plurality of monitored objects; The moving body is a conveying path (5), the sensors include a vibration sensor and a sound sensor; When the status detection unit detects an abnormality using data acquired by the vibration sensor and data acquired by the sound sensor that includes data acquired by the vibration sensor, the status detection unit determines that the abnormality is caused by the conveying path, and when the abnormality is detected using only data acquired by the sound sensor, the status detection unit determines that the abnormality is caused by a factor other than the conveying path.

3. An equipment status monitoring system, comprising: a sensor node (10) that has a sensor (11) that outputs data indicating the state of a facility (2) to be monitored as sensor data, a communication unit (13) that transmits the sensor data, and a power supply unit (12) that supplies power to the sensor and the communication unit, and that is commonly applied to a plurality of the monitored facilities; a receiver (20) that receives the sensor data transmitted from the communication unit; a state detection unit (30) that inputs the sensor data received by the receiver, learns the normal states of the plurality of monitored objects as learning data based on the sensor data in normal times when the plurality of monitored objects are operating normally, and when the sensor data transmitted from the sensor node after learning is received by the receiver, compares the states of the plurality of monitored objects indicated by the sensor data with the learning data to detect the occurrence of an abnormality or a sign of an abnormality in the plurality of monitored objects, a power supply method from the power supply unit to the sensor and the communication unit is one of a method using a magnetic field, energy harvesting, and mobile battery power supply; the sensor node is disposed on a mobile body and moved together with the mobile body to acquire the sensor data indicating the states of the plurality of monitored objects; The moving body is a conveying path (5), The communication unit processes the sensor data in accordance with the conveying speed of the conveying path and transmits the processed sensor data to the receiver.

4. The equipment state monitoring system according to claim 3 , wherein the communication unit is capable of performing downsampling processing on the sensor data, and the slower the conveying speed is, the lower the sampling frequency of the downsampling processing is set.

Images